Development and Implementation of a Remote Audit Tool for High
192
Dose Rate (HDR) Ir Brachytherapy Using Optically Stimulated
Luminescence Dosimetry
Kevin Casey1, Paola Alvarez1, Ann Lawyer1, Stephen Kry1, Rebecca Howell1, Scott Davidson2, David Followill1
of Radiation Physics, UT MD Anderson Cancer Center, Houston, TX, 2The Methodist Hospital, Houston, TX
Introduction
The Radiological Physics Center’s (RPC) mission
is to ensure consistency and comparability of
radiation
doses
delivered
at
institutions
participating in NCI-funded cooperative clinical
trials. A major effort of the RPC to accomplish this
mission is the mailable optically stimulated
luminescence dosimeter (OSLD) program for
remote audits of participating institutions’ external
beam (EBRT) reference calibrations. This
program has accuracy sufficient to establish a
±5% acceptance criterion for comparison
between RPC-measured and institution-reported
dose [1]. However, brachytherapy, which is the
placement of radioactive sources in or near the
tumor, is also used in clinical trials. Unfortunately,
no program analogous to the RPC’s EBRT
program exists for remote audits of high dose-rate
(HDR) brachytherapy sources. Current RPC HDR
activities consist only of plan checks,
questionnaires, and infrequent site visits.
This project aims to create a mailable, OSLDcompatible tool capable of remote audits of HDR
brachytherapy sources with accuracy suitable for
RPC monitoring of clinical trial sites.
Materials (continued)
Methods (continued)
Results (continued)
KB corrects for a number of factors which are
unique to this project, such as OSLD
overresponse at 192Ir spectrum energies,
incomplete backscatter and lack of equivalence
between polystyrene and water, and angular
dependence of OSLD nanoDots. It was
determined by irradiating dosimeters, correcting
the reading for fading and linearity, and then
dividing the TG-43 calculated dose (with a NISTtraceable source strength) at the point of
measurement by the corrected reading. This
isolates KB on one side of Equation 1.
Results
KL - Linearity Correction Factor
1.010
1.010
1.005
1.005
Figure 3: The phantom prototype broken apart
with nanoDots inserted.
Methods
y = -9.433E-05x + 1.009
KL
1Dept
1.000
1.000
The equation used to calculate dose from an
OSLD reading is as follows:
Materials
0.995
0.995
0.990
0.990
(1)
Landauer’s nanoDot OSL dosimeters (Figure 1)
were chosen for their near-planar geometry and
proven accuracy when used as the basis for a
mailed dosimeter program. The RPC has
considerable experience with and infrastructure in
place for using nanoDots, with over 10,000
currently in circulation as part of the external
beam audit program.
Where:
reading
ECF
Sensitivity
KF
KL
KB
raw, uncorrected OSLD reading
element-specific correction factor
system sensitivity at time of
reading [dose/counts]
fading correction factor
linearity correction factor
block/energy correction factor
ECF corrects for differences in sensitivity between
individual dosimeters and the overall batch
average and has already been determined by the
RPC for each dosimeter currently in use.
Figure 1: nanoDot OSL Dosimeters
cm3
An 8 x 8 x 10
phantom prototype was
manufactured out of high-impact polystyrene
(ρ=1.04 g/cm3) (Figure 2). The phantom has a
single channel which admits a standard HDR
endobronchial catheter. Two slots, one on either
side of the channel, hold nanoDot dosimeters. The
phantom breaks into two pieces for ease in
loading and unloading dosimeters (Figure 3).
Sensitivity is determined anew for each OSLD
reading session through the reading of special
“standards” dosimeters. These standards are
dosimeters irradiated to 100 cGy under carefully
controlled conditions using a 60Co beam. Reading
a standard before and after each individual OSLD
reading session allows for the establishment of a
dose-per-OSLD-counts
conversion
(aka
Sensitivity).
KF corrects for OSLD signal fading over time after
irradiation and has been previously established by
the RPC.
KL corrects for the linearity of OSLD response with
dose. It was determined by irradiating 78
dosimeters to doses between 50 and 400 cGy. For
each dosimeter, the nominal dose per OSLD
reading (“dose response”) was plotted against the
nominal dose. A linear fit was applied and
normalized to the value at 100 cGy nominal dose.
Thus, KL  1.000 at 100 cGy.
Figure 2: Cross section of phantom prototype. All
dimensions in mm.
80
85
90
95
100
105
110
115
120
Nominal Dose [cGy]
Figure 4: Linearity correction factor with 95%
confidence interval.
It was found that KL=-9.433×10-5 × Dose +
1.0094 where Dose is the nominal dose in cGy. KL
is shown along with its 95% confidence interval in
Figure 4.
192Ir
KB was determined separately for two
HDR
sources, the Nucletron microSelectron v2 and the
Varian VariSource VS2000. KB results are
summarized in Table 1.
Nucletron
Varian
20
10
Average
1.026
1.000
Standard
Deviation
0.6%
0.7%
n
99% Confidence
1.022 – 1.029 0.993 – 1.007
Interval
Table 1: Block correction factors for two 192Ir HDR
sources.
The percent uncertainty in Dose measurements
using the system was estimated by adding in
quadrature the percent uncertainties of each term
in Equation 1. Uncertainties for reading, ECF,
sensitivity, and KF were provided by the previous
work of Aguirre et al.[2]. Percent uncertainty in KL
was 0.15% in the region of 90-110cGy (see Figure
4). Percent uncertainty in KB was the measured
standard deviation for each source.
Quantity
Uncertainty,
Nucletron
Uncertainty,
Varian
Reading
ECF
Sensitivity
KF
KL
KB
Total (2σ)
0.57
0
0.8
0.3
0.15
0.6
2.4
0.57
0
0.8
0.3
0.15
0.7
2.5
Table 2: Uncertainty budget for dose
measurements.
To date, remote audits have been performed at 8
institutions using the system described here
(Table 3). For each audit, the OSLD-measured
dose was compared to the dose reported by the
institution’s treatment planning computer at the
point of measurement. The average ratio is 1.000
and the standard deviation is 0.011.
Institution
Source Model
1
2
3
4
5
6
7
8
Average
Nucletron
Varian
Varian
Nucletron
Nucletron
Varian
Nucletron
Varian
RPC/Inst
Ratio
0.989
1.005
1.001
0.999
1.014
0.983
1.012
0.996
1.000
Table 3: Institution audit results from system
feasibility study.
Conclusion
The estimated 2σ uncertainty of 2.4% or 2.5% is
sufficient to establish a ±5% acceptance criteria
on RPC-to-institution dose ratios [3].
Furthermore, preliminary remote audit results
compare favorably to a sample of 193 wellchamber measurements performed by the RPC on
site visits from 1994 to 2011. Average RPC-toinstitution ratio for the well-chamber visits was
1.009 with standard deviation of 0.014. This is
compared to 1.000 and 0.011, respectively,
measured with this project.
The tool established in this work is durable, simple,
and most importantly accurate enough for RPC
audits of HDR brachytherapy sources at institutions
participating in NCI-funded clinical trials. This will
greatly help the RPC in pursuit of its mission to
ensure consistent and comparable radiation doses
at these institutions as HDR brachytherapy
becomes ever more prevalent in cooperative
clinical trials.
References
1) Aguirre et al. “WE-D-BRB-08: Validation of the Commissioning of an
Optically Stimulated Luminescence (OSL) System for Remote Dosimetry
Audits” Med Phys 37, 3428 (2010).
2) Aguirre et al. “SU-E-T-126: Analysis of Uncertainties for the RPC Remote
Dosimetry Using Optically Stimulated Light Dosimetry (OSLD)” Med Phys
38, 3515 (2011).
3) Kirby, et al. “Uncertainty analysis of absorbed dose calculations from
thermoluminescence dosimeters” Med Phys 19, 1427-1433 (1992).
Support
This investigation was supported by PHS grant CA10953 awarded by the
NCI, DHHS.